Psoralen plus ultraviolet A (PUVA), commonly used for the treatment of hyperproliferative skin disorders, has been found to be associated with an increased risk of squamous cell cancer. Interstrand cross-link (ICL) formation by PUVA treatment is considered the major factor contributing to the carcinogenesis. However, it remains unclear how PUVA causes, or promotes cancers, in humans. As an initial step in understanding the mechanisms of mutagenesis and carcinogenesis of PUVA photochemotherapy, we have optimized and subsequently utilized a modified alkaline comet assay involving a post-lysis gamma-irradiation at 9 Gy to sensitively measure the formation and repair of PUVA-induced ICLs in the immortalized human keratinocyte cell line HaCaT. A clear dose-dependent response of HaCaT cells to PUVA exposure was observed with a combination of a fixed UVA dose at 0.05 J/cm(2) and a dose of 8-methoxypsoralen ranging from 10 to 100 microM. Results also indicated that the ICL repair was concentration dependent. We have also demonstrated that PUVA-induced monoadduct formation, at an estimated ratio of 3:1 to ICLs in the present experimental conditions, does not interfere with the detection of the ICLs in the modified alkaline comet assay. Furthermore, comparison of the amount of ICL formation between the single-dose UVA treatment and a split-dose protocol was performed. The split-dose protocol was believed to generate more ICLs than the single-dose treatment, thus more effective in PUVA photochemotherapy. Our results demonstrate that comparable amounts of ICLs were formed in HaCaT cells for each dose of UVA, using either the split-dose or single-dose protocols.